[Spacewalk-list] Spacewalk slow opens full list of nodes (Systems) 250nodes

Matt Moldvan matt at moldvan.com
Tue Mar 29 18:34:18 UTC 2016


I was watching load time in Chrome developer tools for 500 systems, took
10.22 seconds to fully load.  Our Postgres tunables look like:

default_statistics_target = 50 # pgtune wizard 2015-12-08
maintenance_work_mem = 960MB # pgtune wizard 2015-12-08
constraint_exclusion = on # pgtune wizard 2015-12-08
checkpoint_completion_target = 0.9 # pgtune wizard 2015-12-08
effective_cache_size = 6GB # pgtune wizard 2015-12-08
work_mem = 3840kB # pgtune wizard 2015-12-08
wal_buffers = 8MB # pgtune wizard 2015-12-08
checkpoint_segments = 256 # pgtune wizard 2015-12-08
shared_buffers = 3840MB # pgtune wizard 2015-12-08
checkpoint_timeout = 30min # range 30s-1h

The above works relatively well, considering we have ~5,000 systems
registered and the Jabber services are also using it for OSAD information.

I don't recall if autovacuum is set to on in the default config, but I also
saw some benefit from running vacuum on a few tables in the Spacewalk
database (rhnserver, rhnerrata, etc).

autovacuum = on # Enable autovacuum subprocess?  'on'
autovacuum_max_workers = 30 # max number of autovacuum subprocesses


On Tue, Mar 29, 2016 at 1:53 PM Konstantin Raskoshnyi <konrasko at gmail.com>
wrote:

> Yep, already did.
>
> default_statistics_target = 100
> maintenance_work_mem = 2GB
> checkpoint_completion_target = 0.9
> effective_cache_size = 44GB
> work_mem = 2048MB
> wal_buffers = 16MB
> checkpoint_segments = 32
> shared_buffers = 14GB
> max_connections = 100
>
> Full load of 250 nodes about 20second. Was 150.
> What's you load time if you can share?
> Thanks!
>
> On Tue, Mar 29, 2016 at 10:44 AM, Matt Moldvan <matt at moldvan.com> wrote:
>
>> Sounds like since you're giving Tomcat more resources it's able to do
>> it's job better, and now Postgres needs some tuning.  I ran pgtune on mine
>> for 2,000 clients as such:
>>
>> as the postgres user:
>> pgtune -i data/postgresql.conf  -o ./data/postgresql.conf.new -c 2000
>>
>> You might be able to get way with fewer, though, but it's worth a try if
>> you want a usable UI without having to update (though the update for me was
>> relatively painless if I recall correctly).
>>
>> On Tue, Mar 29, 2016 at 12:50 PM Konstantin Raskoshnyi <
>> konrasko at gmail.com> wrote:
>>
>>> With this settings JAVA doesn't eat the memory anymore, but postgres eat
>>> 100%.
>>> Loading of the page with this settings is 100second.
>>> With default 70
>>>
>>> On Mon, Mar 28, 2016 at 8:53 PM, Matt Moldvan <matt at moldvan.com> wrote:
>>>
>>>> At those values I'm not surprised you see some slowness... try Xmx/Xms
>>>> of 4G and PermSize of 512m, MaxPermSize of 1024m.  The PermSize I think is
>>>> what helps the most, as the default is only 64MB if it's not specified
>>>> directly.  I'm not a Tomcat expert but it felt faster after I set the
>>>> values that way...
>>>>
>>>> On Mon, Mar 28, 2016 at 10:31 PM Konstantin Raskoshnyi <
>>>> konrasko at gmail.com> wrote:
>>>>
>>>>> I have this values for JAVA_OPTS
>>>>>
>>>>> tomcat   20101  0.3  0.7 4277092 513852 ?      Ssl  Mar25  16:36 java
>>>>> -ea -Xms256m -Xmx256m -Djava.awt.headless=true
>>>>> -Dorg.xml.sax.driver=org.apache.xerces.parsers.SAXParser
>>>>> -Dorg.apache.tomcat.util.http.Parameters.MAX_COUNT=1024 -XX:MaxNewSize=256
>>>>> Tried to increase Xms & Xmx, but didn't find any advantages
>>>>>
>>>>> On Fri, Mar 25, 2016 at 8:09 PM, Matt Moldvan <matt at moldvan.com>
>>>>> wrote:
>>>>>
>>>>>> You mentioned you did some Tomcat tuning, but what are your values
>>>>>> set to?  I noticed some considerable speed up when I tinkered with PermSize
>>>>>> and similar variables, like below.  I think the defaults are pretty low and
>>>>>> not enough memory is allocated off the bat.  Also while the page is loading
>>>>>> what do you notice in top or atop or htop?
>>>>>>
>>>>>> tomcat    2161     1  0 Mar24 ?        00:04:59
>>>>>> /usr/lib/jvm/java/bin/java -XX:NewRatio=4 -XX:PermSize=1024m
>>>>>> -XX:MaxPermSize=2048m -XX:NewSize=2048m -XX:MaxNewSize=2048m -Xms8g -Xmx8g
>>>>>> -XX:+UseParNewGC -XX:+UseConcMarkSweepGC
>>>>>> -Dsun.rmi.dgc.client.gcInterval=3600000
>>>>>> -Dsun.rmi.dgc.server.gcInterval=3600000...rest truncated
>>>>>>
>>>>>> On Fri, Mar 25, 2016 at 9:07 PM Konstantin Raskoshnyi <
>>>>>> konrasko at gmail.com> wrote:
>>>>>>
>>>>>>> Sorry, Asking again...Any possibilities to fix this bug, without
>>>>>>> upgrading the system?
>>>>>>> It's a physical host. Today it's almost not working (Systems menu),
>>>>>>> even if i list 25 hosts or search for a specific one. The top doesn't show
>>>>>>> any processes locking the system.
>>>>>>> Thanks!
>>>>>>>
>>>>>>> On Fri, Mar 25, 2016 at 10:41 AM, Matt Moldvan <matt at moldvan.com>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> Yes I think the way the dependencies are set in the packages
>>>>>>>> updating one will potentially update the rest, so best to do the update off
>>>>>>>> hours, and if possible (if it's a VM for example) take a snapshot or back
>>>>>>>> up of your config and database first.
>>>>>>>>
>>>>>>>> On Fri, Mar 25, 2016 at 12:14 PM Konstantin Raskoshnyi <
>>>>>>>> konrasko at gmail.com> wrote:
>>>>>>>>
>>>>>>>>> Yep. I double checked the version, and it's 2.3.
>>>>>>>>> Is it necessary to update the whole packages with yum?
>>>>>>>>> It's the main deployment server. Don't want to get any troubles
>>>>>>>>> :).
>>>>>>>>> Thanks!
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Friday, March 25, 2016, Matt Moldvan <matt at moldvan.com> wrote:
>>>>>>>>>
>>>>>>>>>> Cool, I was thinking you were already at 2.4 from your earlier
>>>>>>>>>> reply.  Either way good luck and let us know if it solves your issue in
>>>>>>>>>> case anyone has the same question in the future.
>>>>>>>>>>
>>>>>>>>>> > The spacewalk version is 2.4, Oct 7th, 2015 Thanks!
>>>>>>>>>>
>>>>>>>>>> On Thu, Mar 24, 2016 at 11:15 PM Konstantin Raskoshnyi <
>>>>>>>>>> konrasko at gmail.com> wrote:
>>>>>>>>>>
>>>>>>>>>>> Hi Matt!
>>>>>>>>>>>
>>>>>>>>>>> I tried to tune tomcat, but looks like here's my problem
>>>>>>>>>>> https://bugzilla.redhat.com/show_bug.cgi?id=1214437
>>>>>>>>>>>
>>>>>>>>>>> Going to upgrade to version 2.4..
>>>>>>>>>>>
>>>>>>>>>>> Thanks!
>>>>>>>>>>>
>>>>>>>>>>> On Thu, Mar 24, 2016 at 8:06 PM, Matt Moldvan <matt at moldvan.com>
>>>>>>>>>>> wrote:
>>>>>>>>>>>
>>>>>>>>>>>> What are your tomcat settings like?  We have maxThreads set to
>>>>>>>>>>>> 2048 for the 8009 and 8080 connectors in /etc/tomcat6/server.xml:
>>>>>>>>>>>>
>>>>>>>>>>>>     <Connector port="8080" protocol="HTTP/1.1"
>>>>>>>>>>>> connectionTimeout="20000" redirectPort="8443" maxThreads="2048"
>>>>>>>>>>>> maxKeepAliveRequests="1024" URIEncoding="UTF-8" address="127.0.0.1"/>
>>>>>>>>>>>>     <Connector port="8009" protocol="AJP/1.3"
>>>>>>>>>>>> redirectPort="8443" URIEncoding="UTF-8" address="127.0.0.1"
>>>>>>>>>>>> maxThreads="2048" maxConnections="2048" connectionTimeout="600"
>>>>>>>>>>>> keepAliveTimeout="600"/>
>>>>>>>>>>>>
>>>>>>>>>>>> Also, /etc/tomcat6/tomcat6.conf has some settings for JAVA_OPTS
>>>>>>>>>>>> that are interesting for tuning purposes:
>>>>>>>>>>>>
>>>>>>>>>>>> JAVA_OPTS="-XX:NewRatio=4 -XX:PermSize=1024m
>>>>>>>>>>>> -XX:MaxPermSize=2048m -XX:NewSize=2048m -XX:MaxNewSize=2048m -Xms8g -Xmx8g
>>>>>>>>>>>> -XX:+UseParNewGC -XX:+UseConcMarkSweepGC
>>>>>>>>>>>> -Dsun.rmi.dgc.client.gcInterval=3600000
>>>>>>>>>>>> -Dsun.rmi.dgc.server.gcInterval=3600000"
>>>>>>>>>>>>
>>>>>>>>>>>> Listing 500 systems on our master took ~10.22 seconds on an 8
>>>>>>>>>>>> vCPU/32GB RAM VMware VM, with an external Postgres database VM that has 8
>>>>>>>>>>>> vCPU/16GB RAM, so yours could be much quicker with some additional tuning,
>>>>>>>>>>>> I would think.
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> On Thu, Mar 24, 2016 at 10:01 PM Konstantin Raskoshnyi <
>>>>>>>>>>>> konrasko at gmail.com> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>> The spacewalk version is 2.4, Oct 7th, 2015
>>>>>>>>>>>>>
>>>>>>>>>>>>> Thanks!
>>>>>>>>>>>>>
>>>>>>>>>>>>> On Thu, Mar 24, 2016 at 6:21 PM, William H. ten Bensel <
>>>>>>>>>>>>> WHTENBEN at up.com> wrote:
>>>>>>>>>>>>>
>>>>>>>>>>>>>> What version of spacewalk are you running?
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Bill
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> On Mar 24, 2016, at 20:15, Konstantin Raskoshnyi <
>>>>>>>>>>>>>> konrasko at gmail.com> wrote:
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> This email originated from outside of the company. Please use
>>>>>>>>>>>>>> discretion if opening attachments or clicking on links.
>>>>>>>>>>>>>> ------------------------------
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Hi Community!
>>>>>>>>>>>>>> We have a spacewalk server on SCLinux 7.1, Java 1.7 and
>>>>>>>>>>>>>> postgres
>>>>>>>>>>>>>> When I open Systems menu (list servers) and list whole 250
>>>>>>>>>>>>>> nodes - it takes about 90 seconds.
>>>>>>>>>>>>>> The java process shows 400%.
>>>>>>>>>>>>>> The server has 64Gb of Ram and 24 cores.
>>>>>>>>>>>>>> Thought the problem in postgres, actually i got the query
>>>>>>>>>>>>>> from the log file, the query runs about 8seconds.
>>>>>>>>>>>>>> All other time Java doing something...
>>>>>>>>>>>>>> Any solutions?
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thanks!
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> **
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> This email and any attachments may contain information that
>>>>>>>>>>>>>> is confidential and/or privileged for the sole use of the intended
>>>>>>>>>>>>>> recipient. Any use, review, disclosure, copying, distribution or reliance
>>>>>>>>>>>>>> by others, and any forwarding of this email or its contents, without the
>>>>>>>>>>>>>> express permission of the sender is strictly prohibited by law. If you are
>>>>>>>>>>>>>> not the intended recipient, please contact the sender immediately, delete
>>>>>>>>>>>>>> the e-mail and destroy all copies.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> **
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> _______________________________________________
>>>>>>>>>>>>>> Spacewalk-list mailing list
>>>>>>>>>>>>>> Spacewalk-list at redhat.com
>>>>>>>>>>>>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> _______________________________________________
>>>>>>>>>>>>> Spacewalk-list mailing list
>>>>>>>>>>>>> Spacewalk-list at redhat.com
>>>>>>>>>>>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> _______________________________________________
>>>>>>>>>>>> Spacewalk-list mailing list
>>>>>>>>>>>> Spacewalk-list at redhat.com
>>>>>>>>>>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> _______________________________________________
>>>>>>>>>>> Spacewalk-list mailing list
>>>>>>>>>>> Spacewalk-list at redhat.com
>>>>>>>>>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>>>>>>>>>
>>>>>>>>>> _______________________________________________
>>>>>>>>> Spacewalk-list mailing list
>>>>>>>>> Spacewalk-list at redhat.com
>>>>>>>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>>>>>>>
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> Spacewalk-list mailing list
>>>>>>>> Spacewalk-list at redhat.com
>>>>>>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Spacewalk-list mailing list
>>>>>>> Spacewalk-list at redhat.com
>>>>>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Spacewalk-list mailing list
>>>>>> Spacewalk-list at redhat.com
>>>>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Spacewalk-list mailing list
>>>>> Spacewalk-list at redhat.com
>>>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>>>
>>>>
>>>> _______________________________________________
>>>> Spacewalk-list mailing list
>>>> Spacewalk-list at redhat.com
>>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>>>
>>>
>>> _______________________________________________
>>> Spacewalk-list mailing list
>>> Spacewalk-list at redhat.com
>>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>
>>
>> _______________________________________________
>> Spacewalk-list mailing list
>> Spacewalk-list at redhat.com
>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>
>
> _______________________________________________
> Spacewalk-list mailing list
> Spacewalk-list at redhat.com
> https://www.redhat.com/mailman/listinfo/spacewalk-list
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/spacewalk-list/attachments/20160329/45fe648c/attachment.htm>


More information about the Spacewalk-list mailing list